Research On Key Technologies About Segmentation And Enhancement Of Plant Risk Source In Security Inspection Image

Exotic plants will destroy the nature of the original landscape,and there is a risk of destroying the original ecosystem and affecting genetic diversity.However,due to people’s increasingly frequent cross-border exchanges,many plants are easy to break through geographical isolation and flow to other regions.In order to prevent the invasion of foreign plants,it is very necessary to check the entry-exit packages.Although the X-ray machine can carry out fluoroscopic inspection on entry-exit packages,due to the limitation of visual contrast resolution,it is very difficult to find and recognize low-density light and thin objects in X-ray images,which limits the application and development of X-ray machine in entry-exit plant inspection and quarantine.In view of the above problems,this paper designs and implements an image segmentation and local enhancement method based on rough fuzzy set to solve the problem that low-density plant risk sources in entry-exit security inspection X-ray images are difficult to highlight,so as to facilitate the inspection of customs security personnel and improve the customs clearance efficiency of entry-exit passengers and goods.The main work is as follows:(1)According to the characteristics of X-ray plant risk source imaging and the indistinguishable relationship of rough set,a region of interest extraction algorithm for X-ray plant inspection based on rough set subgraph segmentation is introduced.Through X-ray image simulation experiments under different noise levels,the results show that this method can correctly divide the image areas that may contain low-density objects,and provide region of interest data for plant risk source enhancement display processing.(2)Aiming at the region of interest in X-ray plant inspection,a pixel quantitative discrimination method based on fuzzy membership is established,which is used as the basis of pixel enhanced display to realize the local enhanced display of plant risk sources.Through the enhancement of the real X-ray security image containing plant risk sources,the results show that this method can locally enhance the plant risk source area.Compared with other methods,this method is not only better in the effect of local enhancement,but also will not affect other areas of the image.(3)In order to deal with the problem that the contrast is too low and difficult to segment due to too thin plant risk sources,a region of interest growth segmentation method for X-ray plant inspection based on mean shift and adaptive threshold is established on the basis of rough set subgraph segmentation method.The experimental results show that when the contrast of plant risk sources is too low,this method can also reliably extract the region of interest of X-ray plant inspection,and then realize the local enhanced display of plant quarantine objects.